A caching method and system for reducing write performance jitter of a solid-state disk storage system

Abstract

The invention discloses a cache method for reducing write performance jitter of a solid-state disk storage system: S1, when a new write request arrives at the cache system, store the new write request data in a matching cache cluster; S2, based on the new write request data Generate fingerprint information, and match the fingerprint information with the target fingerprint library, and judge the operation type of this new write request according to the matching result; S3, select the new write request data in the new write request according to the operation type of this new write request The processing method when the cache cluster where it is located is written back to the flash memory. The present invention uses the fingerprint library technology to select as many clusters as possible to update the data pages and write them back to the flash memory, and when the storage unit where the original data block is located is idle, garbage collection is performed on the original data block, which improves the utilization rate of the cache space and reduces the cache memory. A large number of invalid data pages generated by frequently updating data and writing back to the flash memory in the medium reduce garbage collection operations and reduce write performance jitter of solid-state disks.

Description

technical field [0001] The invention relates to the technical field of caching optimization methods, in particular to a caching method and system for reducing write performance jitter of a solid-state disk storage system. Background technique [0002] As a new type of storage device, solid state disks have been widely used in various consumer electronic devices as storage media in recent years. With the wide application of flash solid-state disks, the problems caused by the physical characteristics of the storage medium itself are becoming increasingly apparent. The remote update operation brought about by the erase-before-write feature will degrade the performance and lifespan of the solid-state disk. The flash solid-state disk needs to erase the storage space before writing data, so when the data in the solid-state disk needs to be updated, the updated data needs to be written into the data block that has been erased, and then its The corresponding original data is inval...

AI Technical Summary

Problems solved by technology

Garbage collection is mainly composed of two parts: the rewriting operation of valid data pages in the block to be erased and the erasing operation of invalid data blocks. The rewriting operation of valid data pages in the block to be erased may compete with external new write requests for the bus , the external new write request cannot be responded until the rewrite operation is completed, which causes the performance of the external new write request to fluctuate and greatly increases the response time of the user's new write request; Write time, which again increases the time overhead of garbage collection

Larger response delays caused by transactional storage system write performance jitter are not allowed, and large write performance jitter is fatal to transactional storage systems

Images